Open Source in the Share of Wallet Wars

On their blogs, Matt Asay and Savio Rodrigues are dis­cussing whether IBM is using open source to dimin­ish com­peti­tor mar­gins. I think it is obvi­ous that IBM does this, most notably with its Linux engage­ment, which is squarely directed against Microsoft (Win­dows). It is what I call a war over the share of customer’s wal­let, and open source plays a major role. The graph below shows how it works.

By “share of wal­let” I mean the amount (per­cent­age) of a customer’s IT bud­get that a ven­dor can get. Any IT user buys mul­ti­ple dif­fer­ent com­po­nents, for exam­ple, oper­at­ing sys­tem, pro­duc­tiv­ity tools, and busi­ness appli­ca­tions, but ide­ally they only buy one of a given type. They only want one data­base, not many. So each IT ven­dor has two types of com­peti­tors: those that directly com­pete with their prod­ucts, and those that offer ancil­lary prod­ucts. You can take on your direct com­peti­tors through bet­ter sales or ser­vice or pro­duct. For ancil­lary prod­ucts, you need a dif­fer­ent strat­egy.

If you offer a busi­ness appli­ca­tion, data­bases and oper­at­ing sys­tems are ancil­lary (com­ple­men­tary) pro­duct cat­e­gories for you. You aren’t offer­ing a pro­duct in that cat­e­gory, so you can’t make any money of it. You can, how­ever, reduce the amount of money that these other prod­ucts take out of a customer’s IT bud­get (wal­let) and then con­vince the cus­tomer to spend that extra money on your pro­duct. So what you want is to shift money from one com­po­nent in the solu­tion stack to another com­po­nent, specif­i­cally to the one you are offer­ing.

How do you do this? Well, you dimin­ish the mar­gins of the com­po­nents that are ancil­lary to your pro­duct, as Matt right­fully points out. Look at the fig­ure above. If you can make the cus­tomer spend less money on the data­base and oper­at­ing sys­tem, then you can pos­si­bly gain that money for the appli­ca­tion com­po­nent. So, if you are an appli­ca­tion provider, you may want to invest in open source solu­tions for data­base and oper­at­ing sys­tems, if they are cheaper than the lead­ing pro­pri­etary solu­tion (which they typ­i­cally are). This leaves more money on the table for your appli­ca­tion.

I think that’s the open secret behind IBM’s Linux engage­ment. By ensur­ing that there is a viable com­peti­tor to Win­dows, at least on the server, IBM is keep­ing Microsoft’s pric­ing in check, leav­ing more money to be spent on other soft­ware pack­ages. Is there any­thing wrong with it? Not really, it is just com­pet­i­tive strat­egy, to the ben­e­fit of the cus­tomer in this case.

If you want to hear or read more about this, you can turn to an ear­lier arti­cle which dis­cusses the share of wal­let wars or catch me next at Agile 2008, where I will be talk­ing about this too.

6 thoughts on “Open Source in the Share of Wallet Wars

  1. Darren Hague


    This is also inter­est­ing from the imple­men­ta­tion con­sul­tancy point of view. Although the open source com­po­nents are cheaper to “buy”, often they require more resources to imple­ment and sup­port (this is partly because many open source prod­ucts are more focussed on fea­tures than on usabil­ity or doc­u­men­ta­tion). From a con­sul­tancy per­spec­tive, this can be a good thing — even if the oper­at­ing sys­tem or data­base ends up cost­ing them cus­tomer nearly as much as the pro­pri­etary ver­sion, the extra implementation/support money is going to the con­sul­tancy instead of to a soft­ware ven­dor.

  2. Dirk Riehle Post author

    Hi Dar­ren, thanks for the com­ment.

    The con­sult­ing per­spec­tive is actu­ally quite inter­est­ing. One might argue that con­sul­tants would like all soft­ware to be free so that more money is left for con­sult­ing and imple­men­ta­tion ser­vices. How­ever, if every­thing was open source, the bar­ri­ers to entry to the con­sult­ing ser­vices mar­ket would go down sig­nif­i­cantly, as every high-school dropout could set up shop if they only had enough energy to dig into the soft­ware. So you’d want to strike a bal­ance between employ­ing open source and rely­ing on (expen­sive?) pro­pri­etary source that pro­tects your mar­ket value.

  3. Reginald Beauchamps

    It’s a myth that open source prod­ucts are cheaper to buy but have an equal or larger TCO than pro­pri­etary soft­ware. Microsoft have spent a large amount of money sell­ing this myth, so it is not ter­ri­bly sur­pris­ing to find it repeated here.

    Many, many peo­ple are self taught on the linux side of things at sys admin level and less fre­quently at devel­oper level. Nonethe­less the dif­fer­en­tial between rates of pay on the win­dows and unix side are sim­ply con­trolled by sup­ply and demand.

    This was a very inter­est­ing arti­cle. I see the ben­e­fit for IBM post OS/2. Link­ing this to their ODF strat­egy, it seems they have some very decent thinkers in their busi­ness.

  4. Morten Blaabjerg

    Great illus­tra­tion of some­thing which is oth­er­wise hard to illus­trate. I bet your talks will give some­ones some­where some­thing to think about…

    To add to this line of think­ing, I believe this strat­egy can be applied to other types of busi­nesses, espe­cially online, where it is easy to give some­thing away for free. When a com­pany gives some­thing away for free, such as an online ser­vice, it is also fight­ing the “war over the wal­let”.

    Take Google as an exam­ple, which offers a wide range of free online ser­vices, in direct com­pe­ti­tion with “offline” solu­tions such as MSOf­fice and Out­look. By pro­vid­ing pri­vate users as well as busi­nesses ser­vice, they’d oth­er­wise have to pay for, it leaves room for spend­ing that money on, say, online adver­tis­ing, in Google’s case.

  5. James Dixon

    Hi Dirk,

    I agree with your main point and I have been using a dia­gram like this for a few years when I talk to peo­ple about open source and com­mer­cial open source imple­men­ta­tions. I think it is impor­tant to note that the money that is freed up by low­er­ing the cost of the com­po­nents can be spent (and I con­tend usu­ally is spent) on ser­vices not on pro­duct. By spend­ing this money on ser­vices instead of pro­duct you get a more suc­cess­ful out­come for the project.

    If is use­ful at all I have a paper/wiki on the sin­gle ven­dor com­mer­cial open source model called the Bee­keeper:

    James Dixon

  6. Dirk Riehle Post author

    Hi James, this is a smart obser­va­tion, and I agree that more often than not the money is shifted into ser­vices rather than more soft­ware. I think this is one rea­son why IBM is such a strong sup­porter of open source; it plays directly to the strength of its global ser­vices arm.

    I make the argu­ment here using soft­ware because it is the eas­i­est way to explain it. Pre­vi­ously, I used more elab­o­rate eco­nomic mod­els of sup­ply and demand in open source, and they are still valid and prin­ci­pally the same, but then, well, more com­plex to explain.

    How­ever, this added com­plex­ity is not for noth­ing. There are actu­ally mul­ti­ple effects in these eco­nomic games: Less money spent on some soft­ware com­po­nents not only makes it pos­si­ble for cus­tomers to buy more, it even makes it pos­si­ble for ven­dors to reach more price sen­si­tive cus­tomers that were pre­vi­ously out of reach, see the ref­er­enced pub­li­ca­tion above.


Leave a Reply